Parse notes
Anthropic has joined the U.S. Department of Energy's 1,000 Scientist AI Jam, where frontier AI models, including Claude 3.7 Sonnet, will be tested on scientific and national security challenges.
DeepSeek released its inference and revenue numbers at the end of its open source week. The company offers many services for free - if it monetized every token, it would make $200m or so in yearly revenue with healthy margins.
Inception, a Palo Alto startup, has developed a diffusion-based large language model, DLM, which significantly accelerates AI text generation compared to traditional LLMs. Founded by Stanford's Stefano Ermon, the company claims its models are faster, more cost-effective, and offer up to 10x improved performance. It has secured investment from Mayfield Fund and attracted several Fortune 100 customers, offering various deployment options and an API.
A systematic study on reward shaping in RLHF led to Preference As Reward (PAR), a method that uses latent preferences to enhance alignment, improve data efficiency, and mitigate reward hacking, outperforming existing approaches on multiple benchmarks.
A new color space, Horizontal/Vertical-Intensity (HVI), along with the CIDNet model, significantly reduces color artifacts and improves low-light image enhancement.
ReCon introduces a dual-alignment learning framework to improve multimodal correspondence accuracy by enforcing cross-modal and intra-modal relation consistency.
Sponsored
Tech pros scroll past ads all day—but not here. TLDR reaches millions of devs, founders, and decision makers who actually open, read, and engage with what's inside. And to make things easier, TLDR's team writes the ad copy for you, for free. Learn more [10] about running your first ad campaign.
A large-scale empirical study of diffusion planning in offline reinforcement learning.
Through a benchmark using various academic GPUs, this study demonstrates that models like Pythia-1B can be pre-trained in far fewer GPU-days than traditionally required.
VoiceRestore is an advanced tool for restoring and enhancing speech recordings using deep learning aimed at improving clarity and removing noise.
Anthropic tested its Claude 3.7 Sonnet AI model by having it play Pokémon Red, showcasing its "extended thinking" capabilities. The model successfully battled three gym leaders, significantly outperforming its predecessor, Claude 3.0 Sonnet. While details on computing power and time required are unclear, the test highlighted games as a tool for AI benchmarking.
By training an auto-regressive model to generate sheet music, this team has been able to make a novel text-to-music system that is often preferred by human raters.
Perplexity has announced its new browser, Comet, which is aimed at competing with Google's Chrome. No details on its features are available yet. The company has invited users to sign up for beta access in a bid to gain early adopters. Perplexity's move indicates a growing trend of AI-specific applications challenging traditional app categories.
Sam Altman-backed Warp is reimagining the command-line terminal after nearly 40 years unchanged.
Ray now has first-class support for automatic dependence installation based on Python package management uv.
Prime Intellect, a distributed computing company, has raised additional funds to continue pushing its distributed training scheme.